STERILISED MILK. 217 



an equal volume of amyl alcohol the blue substance is dissolved in 

 the alcohol layer, and the test is thus rendered more reliable in the 

 presence of substances which modify the tint (e.g., formaldehyde). 



Other substances, such as quinol or tincture of guaiacum, may 

 be used ; of these the one recommended by Saul, and found most 

 effective by the author, is " ortol," which is sold as a photographic 

 developer. This gives a fine red colour with unboiled milk. 



Benzidine, dissolved in a little acetic acid, as recommended 

 by Wilkinson and Peters, or benzidine hydrochloride, recom- 

 mended by Leffmann, may also be used with advantage. 



These tests show that the milk has been heated above 80 C., 

 as the " peroxydase " which gives rise to the reactions is destroyed 

 at this temperature. 



If the milk has been treated with hydrogen peroxide, the tests 

 with phenylene-diamine or ortol will fail, as the peroxydase is 

 also destroyed if an excess of hydrogen peroxide is added to 

 milk. 



Another test is to add a solution of methylene blue containing 

 formaldehyde and keep the milk out of contact with air ; fresh 

 milk decolourises this, and boiled milk does not. This test 

 appears not to depend on any constituent of the milk, but on 

 the presence of enzymes (reductases) secreted by micro-organisms. 

 Milk treated with hydrogen peroxide gives this test. 



The rate at which methylene blue is decolourised is a rough 

 index of bacterial contamination. 



A cultivated palate may also detect a boiled taste ; this will 

 certainly be noticed with " sterilised " milk. 



It is doubtful whether a proportion of sterilised milk much 

 below 30 per cent, could be detected with certainty when mixed 

 with new milk. 



The proportion of sterilised milk should be deduced from the 

 percentage of soluble albumin by the following formula : 



-4 Percentage of soluble albumin 

 Percentage of sterilised milk = ^ - X 100. 



This is based on the supposition that new milk contains 0-4 per 

 cent, of albumin, while in sterilised milk the albumin has been 

 removed. 



The estimation of albumin is the most reliable test. There 

 are many causes which influence the rising of cream, such as 

 the temperature to which the milk has been warmed or cooled ; 

 the size of the fat globules, which varies with the stage of lacta- 

 tion ; and the acidity of the milk. The milk may also have 

 been " homogenised " by the fat globules having been broken 

 up, in which case practically no cream will rise. The quantitative 

 deductions drawn from observations of the rate of the rising of 



